Presentation template

WS 2025/26

Tim Overkamp

Notes

::: {.notes}
You can add notes here. They will only be visible in the speaker view.
:::

Formatting and custom CSS

Bold text

Italic text

Monospace text

Was ist ein ViT?

[1]

Subpage

This is a subpage

Autoanimate

Test

Autoanimate 2

Test 2
Hallo

vscode extension

Use the quarto extension for vscode to have compile on save and a direct preview.

Quarto Reveal.js docs

click here

Ergebnis

Modell	Accuracy	ECE	MCE
A (Baseline)	0.9738	0.0132	0.1024
A+ (Temp. Scaling)	0.9738	0.0023	0.0163
B (Mod. Targets)	0.9688	0.0234	0.1893
C (Soft Target)	0.9712	0.0321	0.0768
D (MMCE Loss)	0.9729	0.0141	0.1138

Quellen

[1]

A. Dosovitskiy et al., “An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale.” Accessed: Jan. 22, 2026. [Online]. Available: http://arxiv.org/abs/2010.11929